Well my point is that when translating a story element into a game one has to consider whether how the element works in the story is a satisfactory way for it to work in the game. I'm deliberately attempting to compare it to what I think I've read in counter-arguments against the "Well that's not how it works in stories" position, which seems to include the idea that the workings of stories can't be translated into games as they directly appear.
AFAICT, the main division in thought is whether or not plot-protection should work in a game in the way that it works in stories. Or rather, to what degree it should do so. Many, many conversations about effects like fudging, SoD, character death, etc., are really just subsets of that overarching conversation, IMHO.
IMHO, that question cannot be answered without first considering what one wants to achieve. In a game like D&D, I want the possibility of death to be real. What I am looking for is an exploration-based game, where players determine the level of the risk they are willing to undertake, and where greater risk = potentially greater rewards. This colours my opinion on many topics re: fantasy role-playing games.
For instance, in 3e, it is generally advisable that the party all be the same level, as a result of the steep power curve. I believe that this detracts from consequence, and that consequence (along with context) are what make choices meaningful. Likewise the way treasure is managed in 4e (Essentials may address this, though): I want the consequence of missing treasure to remain a possibility. A strong likelihood, even, so that regions of the campaign milieu are not so easily "played out".
Likewise, I want SoD, although in my system a character can choose to lose an action to gain a bonus to that save. (I like resource management, and I want my players to have to make resource management decisions; economy of actions is a resource.)
OTOH, if you run a game in which there are no clues as to what you might be facing, then there is no context to making decisions. Not providing context is as bad -- perhaps far worse -- than not providing consequences. So, no random Medusa in room 14b for me. Nor do I make snakes silent from a distance, or snaky waving hair somehow invisible from farther than 30 feet away. Given a choice between RAW and allowing for interesting choices/consequences.....well, it is the spirit of the game I adhere to. Or, at least, the spirit of the game as I wish to play it!
If I am running a game like Cubicle 7's
Doctor Who, or running a superhero game like
Mutants & Masterminds, SoD is not genre-appropriate, IMHO. It is in conflict with what I want, rather than supporting it. Still, for example, getting hit by a Dalek gun is lethal....as it should be. Players have resources (Action Points) to ameliorate this lethality. In the case of the Player being out of APs, they can be donated by other players (if appropriate), or the GM can offer the Unadventurous trait, which gives the player APs but causes the character to begin thinking about leaving the TARDIS.
Again, I note that those games where the goal is to tell a story are potentially damaged by random losses (whether SoD or otherwise), while those games where the goal is to see what happens are not. Only when you have a vested interest in a particular outcome (or subset of outcomes) does it matter what the actual outcome is.
In a D&D-type game, I don't want the GM to have a vested interest in a particular outcome. Doing so, IMHO, damages player agency.
In a game like
Doctor Who, I expect the GM to have a vested interest in the overarching narrative -- the narrative is (sometimes) more important than player agency in that type of game. You have to know what it is you want in order to know what is appropriate.
OTOH, in a D&D-type game, I want the players to have a vested interest in particular outcomes -- to set goals and to attempt to meet them. The success or failure of the players should be foreordained, IMHO, either because the GM wants to protect the PCs, or because the GM wants to have some kind of gleeful "gotcha" moment. Both of these "undesireable" outcomes arise, you may note, from the GM determining what
should happen, as opposed to accepting what
does happen.
It doesn't matter which side of the screen I am on; that is what I am looking for.
Different strokes for different folks & all that, but I have never wanted for players any time I was willing to run a game. And I have run a lot of games for a lot of people, in several states and two countries. Nor have I ever been in a group where I could be a player for long, as my GMing was in demand. So I can't help but imagine that this type of game, when run well (or at least as well as I can run one) is very popular. At least throughout North America.
YMMV. Different strokes for different folks and all that.
And if YMDoesV, you should play it the way you like it. Life is too short for games you don't enjoy!
RC